Claude Cursor Skill

foxglove

Visualize live or recorded ROS robot data remotely with Foxglove or Lichtblick.

LLM Mart · 0 points · 1 views 0 listing impressions 0 install-command copies
Download robium-ai-robium-skills_foxglove-498ea4e.zip · 5 KB

Install

skills CLI npx skills add https://github.com/robium-ai/robium/tree/main/skills/foxglove
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install robium-ai-robium@llmmart
Git git clone https://github.com/robium-ai/robium.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole robium-ai/robium collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Foxglove and Lichtblick

Treat remote visualization as a chain: ROS graph, bridge, network, client, and layout. Find the first link without evidence.

Choose the client deliberately

  • Lichtblick: prefer an inspectable open-core viewer, local/offline use, or a browser client you can host and extend yourself.
  • Foxglove: prefer its managed data, fleet, organization, and support services, or when the team already uses its layouts and platform.
  • Both clients can consume a Foxglove WebSocket and recordings for the common ROS workflow. Re-check current licensing and self-hosting terms before making them an architecture dependency.

Read client choices when licensing, hosting, extensions, or managed services affect the decision. If the visualization tool itself has not been chosen, use visualization first.

Trace the live-data chain

  • Prove the ROS topics and QoS locally before blaming the bridge.
  • Run the bridge near the ROS graph; DDS discovery should not need to cross a browser, NAT boundary, or public internet.
  • Keep its unauthenticated WebSocket on a private network, VPN, or SSH tunnel.
  • Confirm the listening port and WebSocket reachability before debugging a client layout.
  • Limit high-rate topics at the bridge when bandwidth is constrained. Hiding a panel does not reduce transport load.
  • Save a focused layout per task and record before any run that must be shared, replayed, or compared.

For bridge, MCAP, navigation goals, teleop, actions, layouts, and custom clients, read ROS 2 workflows. For a connection that is blank, unreachable, mixed-content blocked, or missing controls, read failures.

Cross into ros2 when the local graph or QoS is wrong. Cross into environments when the bridge host, VPN, or container cannot provide the network contract. Cross into navigation only after pose, map, sensors, and goal messages are proven across the viewer boundary.

Done

  • The intended client reaches the bridge without exposing it publicly.
  • The necessary topics render at useful rates and with correct frames.
  • Controls publish to the interfaces the robot actually consumes.
  • The layout and recording path can be reused by another person.
  • Current behavior and hosting terms were checked against official Foxglove and Lichtblick documentation.
Files (robium)
  • CLIENTS.md 2 KB
    Loading…
  • evals.yaml 437 B
    Loading…
  • FAILURES.md 2.2 KB
    Loading…
  • ROS2-WORKFLOWS.md 2.7 KB
    Loading…
  • SKILL.md 2.5 KB
    ---
    name: foxglove
    description: Visualize live or recorded ROS robot data remotely with Foxglove or Lichtblick.
    ---
    
    # Foxglove and Lichtblick
    
    Treat remote visualization as a chain: ROS graph, bridge, network, client, and
    layout. Find the first link without evidence.
    
    ## Choose the client deliberately
    
    - **Lichtblick:** prefer an inspectable open-core viewer, local/offline use, or
      a browser client you can host and extend yourself.
    - **Foxglove:** prefer its managed data, fleet, organization, and support
      services, or when the team already uses its layouts and platform.
    - Both clients can consume a Foxglove WebSocket and recordings for the common
      ROS workflow. Re-check current licensing and self-hosting terms before making
      them an architecture dependency.
    
    Read [client choices](CLIENTS.md) when licensing, hosting, extensions, or
    managed services affect the decision. If the visualization tool itself has not
    been chosen, use `visualization` first.
    
    ## Trace the live-data chain
    
    - Prove the ROS topics and QoS locally before blaming the bridge.
    - Run the bridge near the ROS graph; DDS discovery should not need to cross a
      browser, NAT boundary, or public internet.
    - Keep its unauthenticated WebSocket on a private network, VPN, or SSH tunnel.
    - Confirm the listening port and WebSocket reachability before debugging a
      client layout.
    - Limit high-rate topics at the bridge when bandwidth is constrained. Hiding a
      panel does not reduce transport load.
    - Save a focused layout per task and record before any run that must be shared,
      replayed, or compared.
    
    For bridge, MCAP, navigation goals, teleop, actions, layouts, and custom clients,
    read [ROS 2 workflows](ROS2-WORKFLOWS.md). For a connection that is blank,
    unreachable, mixed-content blocked, or missing controls, read
    [failures](FAILURES.md).
    
    Cross into `ros2` when the local graph or QoS is wrong. Cross into
    `environments` when the bridge host, VPN, or container cannot provide the
    network contract. Cross into `navigation` only after pose, map, sensors, and
    goal messages are proven across the viewer boundary.
    
    ## Done
    
    - The intended client reaches the bridge without exposing it publicly.
    - The necessary topics render at useful rates and with correct frames.
    - Controls publish to the interfaces the robot actually consumes.
    - The layout and recording path can be reused by another person.
    - Current behavior and hosting terms were checked against official
      [Foxglove](https://docs.foxglove.dev/docs) and
      [Lichtblick](https://lichtblick-suite.github.io/docs/) documentation.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.